Motion Vector Predictor Generation for Video Encoding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video encoding and decoding methods face inefficiencies in predicting motion vectors, particularly when adjacent blocks do not share the same reference picture, leading to suboptimal compression efficiency and increased data transmission costs.

Innovation Solution

The method generates a motion vector predictor by using either the motion vector of an adjacent block referring to the same reference picture or an alternative reference picture, depending on availability, and encodes the difference between this predictor and the current block's motion vector, allowing for adaptive prediction based on adjacent block characteristics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If motion vector prediction uses adjacent blocks with different reference pictures, then prediction can be performed in more cases, but prediction accuracy deteriorates

Engineering Contradiction:

Improveprediction availabilityVSAvoidprediction accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent applies local quality by differentiating the prediction process based on reference picture matching. When adjacent blocks reference the same picture, their motion vectors are used directly. When reference pictures differ, temporal scaling is applied locally to adjust the motion vector according to the specific reference picture timing, ensuring prediction accuracy is maintained in each local case

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent changes the motion vector parameters through temporal scaling when reference pictures differ. The motion vector from adjacent blocks is scaled by a factor derived from the temporal distances between current picture, adjacent block's reference picture, and current block's reference picture, adapting the parameter to match the current prediction context

Inventive Principle:

Principle #35Parameter changes

2Reliability

If motion vector prediction uses median of multiple adjacent blocks, then prediction robustness improves, but encoding complexity increases

Engineering Contradiction:

Improveprediction robustnessVSAvoidencoding complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies partial action by selectively using motion vectors from adjacent blocks based on reference picture matching. Instead of always computing the median of all adjacent blocks, it uses the motion vector directly when reference pictures match, and only applies temporal scaling when needed, reducing unnecessary computational operations

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent segments the prediction process into distinct cases: when adjacent blocks reference the same picture and when they reference different pictures. This segmentation allows different prediction strategies to be applied to different segments, simplifying the overall process compared to always using median computation

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS9451268B2Methods and apparatuses for encoding and decoding motion vector

Publication Date: 2016.09.20 SAMSUNG ELECTRONICS CO LTD

AI summary

Encoding and decoding a motion vector using a motion vector of a current block of a current picture, which indicates a region corresponding to the current block in a first reference picture and one of generating a motion vector predictor from a motion vector of the adjacent block having a motion vector referring to the first reference picture among adjacent blocks encoded before the current block and a motion vector of an adjacent block referring to a second reference picture other than the first reference picture.